Since when do the R500 support SM3.0?

[KimB]

Doesn't require it, but then neither does dynamic branches. The question is how fast you want it to be...

But then the implementation details of static branches become pretty much identical to those for dynamic branches, and thus there is no meaning for static branches unless you talk about them as being unrolled by the driver.

 

[Hellbinder]

R420 was defeloped by the Marlbro team..

R520 was/is developed by the California team.

While the Core technology is the same or based on the same. The 520 has been in Development since the R400 Fiasco 18ish months ago, And was taking place Simultaniously to the development and release of the R420 by the East Coast team.

In other Words, Dont expect the R520 to be an R420 with SM3.0+ "tacked" on.

 

[Dave Baumann]

R420 was defeloped by the Marlbro team..

CA were the primary driving force behing that.

 

[Acert93]

The 520 has been in Development since the R400 Fiasco 18ish months ago...

I am not an insider. Can I ask what fiasco? All I had heard was from small statements (maybe from Dave?) that the R400 was feature rich and too performance poor to push those features. ATI scrapped the R400 uised the research for the Xbox R500 and the forthcoming PC R600. Is that the fiasco, or was there more/something else?

Also, it is interesting that if R400 got shelved that ATI had the R420 and R520 waiting in the wings. Is it common to have two designs competing for the same time table? It would seem the effort would be best to have everyone focused on one project per timetable instead of two. Maybe the R400 was purely a research project? Make an internal product, see how the actual silicon performs, then redo it and correct the weaknesses. There would be a lot less surprises doing something like this. While expensive, this would create a more mature product that could compete better on the performance front. Possibilities abound... the simplest answer is probably the right one though.

I am interested in knowing more

 

[Dave Baumann]

First carefully read this interview with Dave Orton and then this new post following the interview. They are good places to start.

 

[Acert93]

First carefully read this interview with Dave Orton and then this new post following the interview. They are good places to start.

Thank you very much Dave

Btw, I would like to thank the B3D staff (Reverand, Sonic, JVD, yourself, any others I have not "met") and the community here for all their thoughts and insights. There are a lot of thoughtful and knowledgable people here and it makes it a fun place to read, ask questions, and participate. Thanks!

 

[Dave Baumann]

We're always pleased to see people participate in a manner which befits the site.

 

[Reverend]

We're always pleased to see people participate in a manner which befits the site.

 

[Unknown Soldier]

Richard recently gave me cause to believe that it won't necessarily be unified with that pool (but that doesn't preclude utilising the results before and after).

You giving "other" people food for thought again Dave.

Rev .. you really suprised this site can behave?

US